Hi, In our study we are going to evaluate the effect of 3 different treatments on burn wound using rat models. what should the P-value cutoff be For statistical analysis of our q-pcr output and collagen volume fraction results from image j?(0.01 or 0.05)
Currently statistical researchers suggest reducing error rates with studies that can be replicated with a margin of .005
Hello Sepideh,
If it were a matter of one size truly fitting all cases, then you obviously wouldn't have to ask this question!
The answer to the appropriate level of risk for type I ("alpha level") as well as the acceptable level of risk for type II ("beta level") errors in hypothesis testing really should depend on the consequences of being wrong.
One consequence that large scale replication projects have shown is that a lot of published studies that showed "statistical significance" in the original study fail to do so when a replication is attempted. That's part of the reasoning underlying Cristian Ramos-Vera 's suggestion of a more conservative alpha level. Is it possible to quantify the costs associated with a Type I and Type II error in the framework of your proposed study? That might help you to set realistic levels.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis